Turning on the AI assistant

Effective date: 18 October 2026.

The assistant is an AI, not a person. Its replies are generated by Claude, a model made by Anthropic, PBC. They can be wrong and are not medical, psychological, legal or financial advice. In an emergency, contact local emergency services.

By turning the assistant on, I agree that Alexey Grosman, individual entrepreneur (OGRNIP 320253600063628, INN 250808746453, Moscow, Russian Federation, gros049494@yandex.ru), sends my data outside Russia so the assistant can reply to me.

Recipients:

  • Anthropic, PBC, 548 Market Street, PMB 90375, San Francisco, CA 94104, USA — generates the assistant's replies with its Claude model;
  • Cloudflare, Inc., 101 Townsend Street, San Francisco, CA 94107, USA — relays requests to Anthropic and transcribes voice messages (the audio is not stored, only the text).

The USA does not have an adequacy decision under Russian law. Transfers to these companies are covered by their data processing terms, including the European Commission's Standard Contractual Clauses. US authorities may still be able to request access to data under US law.

What is sent — only what's needed for a particular reply:

  • my message to the assistant or the transcript of my voice message;
  • display name, gender, time zone, chosen life areas;
  • goals, habits, quests and progress, my answers from getting to know the System, confirmed entries in "What the System remembers";
  • today's journal entry and my mood over the last few days;
  • the latest messages in our chat;
  • weekly averages of steps and workouts, if I connected them.

Not sent: Telegram ID, username, email address, profile photo, payment details.

What the recipients do with it. Under its commercial terms Anthropic does not use this data to train its models and keeps it only for a limited time to prevent abuse. Cloudflare does not use requests for training.

Duration and turning it off. Until I turn the assistant off or delete my account. I can turn it off any time in Settings, by messaging the bot or by emailing gros049494@yandex.ru — sending stops immediately and the assistant switches to built-in replies. Data sent before that is deleted by the recipients under their own rules.

The app works without the assistant: it just answers with built-in replies.